Re:Time Series Forecasting with Graph Transformers
by loehnsberg on Tuesday June 17, 2025 @03:09PM
(#44302725)
I disregard anything on time series forecasting from any entity that uses Facebook Prophet as a benchmark.
[ Parent
| on HN ]
Re:Time Series Forecasting with Graph Transformers
by frakt0x90 on Tuesday June 17, 2025 @03:22PM
(#44302864)
Prophet is great and we use it for multiple models in production at work. Our industry has tons of weird holidays and seasonality and prophet handles that extremely well.
Re:Time Series Forecasting with Graph Transformers
by drewbitt on Tuesday June 17, 2025 @07:43PM
(#44305177)
We also used it at my previous job. Yes it does handle that well, but it was also simply not as correct as we would have liked (often over adjusting based on seasonality) even with tuning. Prophet was probably the right choice initially though just on how easy it is to set up to get decent results.
Re:Time Series Forecasting with Graph Transformers
by tech_ken on Tuesday June 17, 2025 @03:45PM
(#44303097)
This is sales research, and after "CAGR in a GSheet" FB Prophet is what's going to be most recognizable to the widest base of customers.
FWIW seems like the real value add is this relational DB model: https://kumo.ai/research/relational-deep-learning-rdl/ The time-series stuff is them just elaborating the basic model structure a little more to account for time-dependence
Re:Time Series Forecasting with Graph Transformers
by melenaboija on Tuesday June 17, 2025 @03:19PM
(#44302835)
For such strong and personal statement I have to ask why.
Re:Time Series Forecasting with Graph Transformers
by Worksheet on Tuesday June 17, 2025 @04:54PM
(#44303846)
If you arrived into, say, London and googled "Best fish and chips" would you believe that the top result gives you the meal that you're after?
Re:Time Series Forecasting with Graph Transformers
by motoxpro on Tuesday June 17, 2025 @07:47PM
(#44305199)
I would believe those are some of the better options and definitely a useful benchmark.
1. How do you go about finding the "absolute" best when you go to a city
2. What does this have to do with the GP's question?
Re:Time Series Forecasting with Graph Transformers
by hotstickyballs on Tuesday June 17, 2025 @06:24PM
(#44304635)
Why not? It’s definitely a useful benchmark
Re:Time Series Forecasting with Graph Transformers
by gk1 on Tuesday June 17, 2025 @08:33PM
(#44305505)
…yes? Feels like there’s some bit of tribal knowledge required to understand your point, but fewer people know it than you think.
Re:Time Series Forecasting with Graph Transformers
by randomcarbloke on Wednesday June 18, 2025 @05:04AM
(#44308051)
as a Londoner I want to urge you to rethink your position.
Re:Time Series Forecasting with Graph Transformers
by esafak on Tuesday June 17, 2025 @03:17PM
(#44302800)
Why? That is what everybody uses. What do you use?
Re:Time Series Forecasting with Graph Transformers
by loehnsberg on Tuesday June 17, 2025 @04:55PM
(#44303852)
L1-regularized autoregressive features, holiday dummies, Fourier terms (if suitable in combination) yield lower test errors, are faster in training, and easier to cross-validate than Prophet.
Re:Time Series Forecasting with Graph Transformers
by hotstickyballs on Tuesday June 17, 2025 @06:24PM
(#44304640)
Sounds like prophet with extra steps